import tifffile 
from matplotlib import pyplot as plt
from skimage import (exposure, feature, filters, io, measure,
                      morphology, restoration, segmentation, transform,
                      util)
import numpy as np
from skimage.morphology import remove_small_objects, dilation, erosion, ball  
from skimage.feature import peak_local_max
from scipy import ndimage as ndi
from skimage.measure import label
import cv2
#from aicssegmentation.core.seg_dot import dot_3d, dot_3d_wrapper 
#from aicssegmentation.core.pre_processing_utils import intensity_normalization, image_smoothing_gaussian_slice_by_slice
import skimage.io
from skimage.measure import label, regionprops, regionprops_table
#import skimage.measure
import pandas as pd
from pathlib import Path
import glob
import seaborn as sns
import nd2
from skimage import img_as_uint
from skimage.util import invert
from vedo import Volume
import pytrax as pt
import pyvista as pv
import os

os.chdir('G:\\D\\AMR_PHD_PROJECT\\OUT\\')
rootdir = Path('G:\\D\\AMR_PHD_PROJECT\\IMAGES')
file_list = [f for f in rootdir.resolve().glob('**/*') if f.is_file()]
file_list=pd.DataFrame(data=file_list)
file_list.columns = ['Data_Path']
file_list = file_list.applymap(str)
file_list['TEXT_1'] = file_list['Data_Path'].str.lstrip(r"G:\\D\AMR_PHD_PROJECT\IMAGES\Mouse\*")
file_list=file_list.assign(IMAGE_ID=file_list['TEXT_1'].str.extract('(.*)M'))
file_list['IMAGE_ID'] = file_list['IMAGE_ID'].str.replace('\\', '')
file_list['IMAGE_ID'] = "MS" + file_list['IMAGE_ID'] 
file_list=file_list.dropna()
UNIQUE_ID =np.unique(file_list['IMAGE_ID'].astype(str))
out='G:\\D\\AMR_PHD_PROJECT\\OUT\\'
ref_df = pd.read_csv('G:\\D\\AMR_PHD_PROJECT\\sod1_mouse_ref.csv')
#141-ROI3 eg of image name
ref_df['ID'] = ref_df['Folder name'].astype(str)
ref_df['IMAGE_ID'] = ref_df['Folder name']

Cell_Volume = []
Void_Volume = []
Protein_in_Cell_Volume = []
Protein_in_Void_Volume = []

for ID in UNIQUE_ID:
    ID_BOOL=file_list.IMAGE_ID ==  ID
    SUB_DF = file_list[ID_BOOL]
    ID_TXT= ID
    
    ID_BOOL_ref=ref_df.IMAGE_ID ==  ID
    ref_sub = ref_df[ID_BOOL_ref]
    Genotype_id=ref_sub['Genotype']
    Genotype_id_string=str(np.array(Genotype_id))
    
    #Read in Cells
    cell_df = SUB_DF[SUB_DF['TEXT_1'].str.contains("cells")]
    cell_path=(cell_df.Data_Path)
    cell_path=np.array(cell_path)
    
    cell_img=tifffile.imread(cell_path[0])
    print("Images from folder")
    print(cell_df.IMAGE_ID)
    print("shape of image Z,X, Y:" )
    print(cell_img.shape)
    print("Number of unique cells:")
    print(np.unique(cell_img))
    
    nd2_df=SUB_DF[SUB_DF['TEXT_1'].str.contains("nd2")]
    nd2_path=(nd2_df.Data_Path)
    nd2_path=np.array(nd2_path)
    nd2_img=nd2.ND2File(nd2_path[0])
    
    #extract voxel size from nd2 raw file
    voxel_size=nd2_img.voxel_size()
    X_VOX_SIZE =voxel_size[0]
    Y_VOX_SIZE =voxel_size[1]
    Z_VOX_SIZE =voxel_size[2]
    print(nd2_img.voxel_size())
    
    #Read in protein 
    protein_df = SUB_DF[SUB_DF['TEXT_1'].str.contains("protein")]
    protein_path=(protein_df.Data_Path)
    protein_path=np.array(protein_path)
    protein=tifffile.imread(protein_path[0])

    #Cells are segmented indivdually, here we make them binary
    cells_binary=cell_img.copy()
    cells_binary[cells_binary !=0 ] = 1
    
    #Void is any "0" pixel in the cell 3D stack
    void_binary=cell_img.copy()
    void_binary = (void_binary == 0).astype(np.uint8)
    
    # using "*" essentially locks the object to its source, so only the protein in cells or void below
    protein_in_cell = cells_binary * protein
    protein_in_void = void_binary * protein
    
    #Create STL's using Vedo and Plot using Pyvista
    CELL_VOL = Volume(cells_binary, spacing=(Z_VOX_SIZE,Y_VOX_SIZE,X_VOX_SIZE))
    CELL_ISO = CELL_VOL.isosurface(1)
    CELL_ISO.write(ID+"CELLS.stl")
    CELL_VOL_STL=pv.read(ID+"CELLS.stl")
    
    PROTEIN_VOL = Volume(protein, spacing=(Z_VOX_SIZE,Y_VOX_SIZE,X_VOX_SIZE))
    PROTEIN_ISO =  PROTEIN_VOL.isosurface(1)
    PROTEIN_ISO.write(ID+"ALL_PROTEIN.stl")
    PROTEIN_VOL_STL=pv.read(ID+"ALL_PROTEIN.stl")

    
    CELL_PROTEIN_VOL = Volume(protein_in_cell, spacing=(Z_VOX_SIZE,Y_VOX_SIZE,X_VOX_SIZE))
    CELL_PROTEIN_ISO = CELL_PROTEIN_VOL.isosurface(1)
    CELL_PROTEIN_ISO.write(ID+"CELL_PROTEIN.stl")
    CELL_PROTEIN_STL=pv.read(ID+"CELL_PROTEIN.stl")
    
    VOID_PROTEIN_VOL = Volume(protein_in_void, spacing=(Z_VOX_SIZE,Y_VOX_SIZE,X_VOX_SIZE))
    VOID_PROTEIN_ISO = VOID_PROTEIN_VOL.isosurface(1)
    VOID_PROTEIN_ISO.write(ID+"VOID_PROTEIN.stl")
    VOID_PROTEIN_STL=pv.read(ID+"VOID_PROTEIN.stl")
    

    #Need an if else statement for when there is no protein 
    if PROTEIN_VOL_STL.n_points != 0:
        pv.set_plot_theme("document")
        p = pv.Plotter()
        p.add_mesh(CELL_VOL_STL, color="grey", opacity=0.4).rotate_y(90);
        p.add_title(Genotype_id_string)
        p.add_mesh(PROTEIN_VOL_STL, color="red", opacity=1).rotate_y(90);
        p.window_size = 500, 500
        p.show(jupyter_backend='static');
        p.screenshot(ID+Genotype_id_string+'3D_VIS.png') ;
        
    else:
        pv.set_plot_theme("document")
        p = pv.Plotter()
        p.add_mesh(CELL_VOL_STL, color="grey", opacity=0.4).rotate_y(90);
        p.add_title(Genotype_id_string)
        p.window_size = 500, 500
        p.show(jupyter_backend='static')
        p.screenshot(ID+Genotype_id_string+'3D_VIS.png') ;
    

    #use regions props to extract volume (um^3)
    Protein_Labels = label((protein_in_cell))
    PROTEIN_INFO = skimage.measure.regionprops_table(Protein_Labels, intensity_image=cell_img,  properties = ('label','area', 'intensity_max'))
    PROTEIN_INFO=pd.DataFrame(PROTEIN_INFO)
    PROTEIN_INFO.columns = ['Protein_Label', 'Protein_Area_Pixels', "Cell_Label"]
    PROTEIN_INFO['Proten_Volume_Voxel_Resolved'] = PROTEIN_INFO['Protein_Area_Pixels'] * X_VOX_SIZE *Y_VOX_SIZE*Z_VOX_SIZE
    PROTEIN_INFO["ID"] = ID
    Protein_in_Cell_Volume.append(PROTEIN_INFO)
    print("Protein Volume Inside Cells'")
    print(PROTEIN_INFO['Proten_Volume_Voxel_Resolved'])
    

    Protein_Labels_void = label((protein_in_void))
    PROTEIN_INFO_void = skimage.measure.regionprops_table(Protein_Labels_void, properties = ('label','area'))
    PROTEIN_INFO_void=pd.DataFrame(PROTEIN_INFO_void)
    PROTEIN_INFO_void.columns = ['Protein_Label', 'Protein_Area_Pixels']
    PROTEIN_INFO_void['Proten_Volume_Voxel_Resolved'] = PROTEIN_INFO_void['Protein_Area_Pixels'] * X_VOX_SIZE *Y_VOX_SIZE*Z_VOX_SIZE
    PROTEIN_INFO_void["ID"] = ID
    Protein_in_Void_Volume.append(PROTEIN_INFO_void)
    print("Protein Volume Inside Void'")
    print(PROTEIN_INFO_void['Proten_Volume_Voxel_Resolved'])
    
    
    CELL_INFO = skimage.measure.regionprops_table(cell_img,  properties = ('label','area'))
    CELL_INFO=pd.DataFrame(CELL_INFO)
    CELL_INFO.columns = ['Cell_Label', 'Cell_Area_Pixels', ]
    CELL_INFO['Cell_Volume_Voxel_Resolved'] = CELL_INFO['Cell_Area_Pixels'] * X_VOX_SIZE *Y_VOX_SIZE*Z_VOX_SIZE
    CELL_INFO["ID"] = ID
    Cell_Volume.append(CELL_INFO)
    
    print("Cell Volume")
    print(CELL_INFO['Cell_Volume_Voxel_Resolved'])
    
    VOID_INFO = skimage.measure.regionprops_table(void_binary,  properties = ('label','area'))
    VOID_INFO=pd.DataFrame(VOID_INFO)
    VOID_INFO.columns = ['Void_Label', 'Void_Area_Pixels', ]
    VOID_INFO['Void_Volume_Voxel_Resolved'] = VOID_INFO['Void_Area_Pixels'] * X_VOX_SIZE *Y_VOX_SIZE*Z_VOX_SIZE
    VOID_INFO["ID"] = ID
    Void_Volume.append(VOID_INFO)
    
    print("Void Volume")
    print(VOID_INFO['Void_Volume_Voxel_Resolved'])
    
    
    print("Move On to Next Image")

print("No more images left")

Images from folder
1    MS100-1_Python
Name: IMAGE_ID, dtype: object
shape of image Z,X, Y:
(16, 1024, 1024)
Number of unique cells:
[ 0  2  3  4  5  6  7  8  9 10 12 13 14 15 16 17 18 19]
VoxelSize(x=0.1022364652365585, y=0.1022364652365585, z=0.3)

Protein Volume Inside Cells'
0    0.062714
1    0.021950
2    0.012543
3    0.116020
4    0.144242
5    0.018814
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Protein Volume Inside Void'
0    0.078392
1    0.003136
2    0.081528
3    0.003136
4    0.012543
5    0.018814
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Cell Volume
0      356.458791
1      806.499069
2      411.863270
3      587.697000
4      882.950289
5     1218.346661
6      225.791518
7      841.995062
8      151.572908
9      402.192807
10     646.873713
11     928.659219
12     447.732411
13     451.429387
14     412.286588
15     386.984718
16     130.877364
Name: Cell_Volume_Voxel_Resolved, dtype: float64
Void Volume
0    43317.911612
Name: Void_Volume_Voxel_Resolved, dtype: float64
Move On to Next Image
Images from folder
4    MS100-6_Python
Name: IMAGE_ID, dtype: object
shape of image Z,X, Y:
(15, 1024, 1024)
Number of unique cells:
[ 0  2  3  4  5  6  7  8  9 11 12 13 14 15 16 17 18 19 20]
VoxelSize(x=0.1022364652365585, y=0.1022364652365585, z=0.3)

C:\Users\richa\AppData\Local\Temp\ipykernel_22212\412934641.py:27: UserWarning: ND2File file not closed before garbage collection. Please use `with ND2File(...):` context or call `.close()`.
  nd2_img=nd2.ND2File(nd2_path[0])

Protein Volume Inside Cells'
0    2.489737
1    0.025086
2    0.006271
3    0.047035
4    0.015678
5    0.094071
6    0.037628
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Protein Volume Inside Void'
0     0.043900
1     0.003136
2     0.003136
3     0.006271
4     0.006271
5     0.003136
6     0.003136
7     0.003136
8     0.003136
9     0.003136
10    0.003136
11    0.006271
12    0.003136
13    0.003136
14    0.003136
15    0.006271
16    0.003136
17    0.053307
18    0.003136
19    0.018814
20    0.015678
21    0.003136
22    0.003136
23    0.025086
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Cell Volume
0     447.475284
1     747.140486
2     691.980591
3     375.376400
4     484.899726
5     526.529126
6     310.085095
7     654.534199
8     416.021193
9      67.984861
10    182.688345
11    110.846587
12    123.113400
13    172.616513
14    242.363631
15    949.633839
16    440.821353
17    182.813772
Name: Cell_Volume_Voxel_Resolved, dtype: float64
Void Volume
0    42193.190337
Name: Void_Volume_Voxel_Resolved, dtype: float64
Move On to Next Image
Images from folder
7    MS101-6_Python
Name: IMAGE_ID, dtype: object
shape of image Z,X, Y:
(18, 1024, 1024)
Number of unique cells:
[ 0  2  3  4  5  6  7  8  9 10 11 12 13]
VoxelSize(x=0.1022364652365585, y=0.1022364652365585, z=0.3)

C:\Users\richa\AppData\Local\Temp\ipykernel_22212\412934641.py:27: UserWarning: ND2File file not closed before garbage collection. Please use `with ND2File(...):` context or call `.close()`.
  nd2_img=nd2.ND2File(nd2_path[0])

Protein Volume Inside Cells'
0    0.009407
1    0.125428
2    0.279076
3    0.109749
4    0.122292
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Protein Volume Inside Void'
0     0.025086
1     0.003136
2     0.012543
3     0.006271
4     0.159920
5     0.003136
6     0.003136
7     0.015678
8     0.047035
9     0.159920
10    0.147377
11    0.009407
12    0.025086
13    0.053307
14    0.003136
15    0.003136
16    0.006271
17    0.169327
18    0.003136
19    0.021950
20    0.464082
21    0.021950
22    1.567844
23    0.003136
24    0.043900
25    0.009407
26    0.116020
27    0.025086
28    0.003136
29    0.003136
30    0.012543
31    0.015678
32    0.018814
33    0.015678
34    0.018814
35    0.012543
36    0.025086
37    0.012543
38    0.009407
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Cell Volume
0      489.324182
1      889.782954
2      859.394997
3     1110.435078
4     1206.145697
5      884.420927
6      704.930984
7      118.980562
8      182.562917
9      614.243739
10     749.156734
11    1039.505806
Name: Cell_Volume_Voxel_Resolved, dtype: float64
Void Volume
0    50335.253109
Name: Void_Volume_Voxel_Resolved, dtype: float64
Move On to Next Image
Images from folder
11    MS101-7_Python
Name: IMAGE_ID, dtype: object
shape of image Z,X, Y:
(18, 1024, 1024)
Number of unique cells:
[ 0  2  3  4  5  6  7  8  9 10 11 12]
VoxelSize(x=0.1022364652365585, y=0.1022364652365585, z=0.3)

C:\Users\richa\AppData\Local\Temp\ipykernel_22212\412934641.py:27: UserWarning: ND2File file not closed before garbage collection. Please use `with ND2File(...):` context or call `.close()`.
  nd2_img=nd2.ND2File(nd2_path[0])

Protein Volume Inside Cells'
0    0.037628
1    0.059578
2    0.028221
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Protein Volume Inside Void'
0     0.072121
1     0.003136
2     0.003136
3     0.106613
4     0.003136
5     0.059578
6     0.059578
7     0.006271
8     0.003136
9     0.003136
10    0.147377
11    0.006271
12    0.153649
13    0.028221
14    0.025086
15    0.219498
16    0.075257
17    0.247719
18    0.181870
19    0.100342
20    0.059578
21    0.012543
22    0.012543
23    0.075257
24    0.056442
25    0.075257
26    0.025086
27    0.009407
28    0.006271
29    0.006271
30    0.003136
31    0.006271
32    0.015678
Name: Proten_Volume_Voxel_Resolved, dtype: float64
Cell Volume
0      955.748432
1     1283.205241
2     1053.619539
3     1123.796247
4     1018.575085
5      662.147651
6     1728.360115
7      176.451460
8      858.767859
9      435.547125
10     219.896424
Name: Cell_Volume_Voxel_Resolved, dtype: float64
Void Volume
0    49668.022507
Name: Void_Volume_Voxel_Resolved, dtype: float64
Move On to Next Image
Images from folder
15    MS102-1_Python
Name: IMAGE_ID, dtype: object
shape of image Z,X, Y:
(15, 1024, 1024)
Number of unique cells:
[ 0  2  3  4  5  6  7  8  9 10 11 12 13]
VoxelSize(x=0.1022364652365585, y=0.1022364652365585, z=0.3)

C:\Users\richa\AppData\Local\Temp\ipykernel_22212\412934641.py:27: UserWarning: ND2File file not closed before garbage collection. Please use `with ND2File(...):` context or call `.close()`.
  nd2_img=nd2.ND2File(nd2_path[0])

Protein Volume Inside Cells'
Series([], Name: Proten_Volume_Voxel_Resolved, dtype: float64)
Protein Volume Inside Void'
0     0.357468
1     0.263398
2     0.056442
3     0.103478
4     0.021950
        ...   
63    0.006271
64    0.003136
65    0.003136
66    0.003136
67    0.003136
Name: Proten_Volume_Voxel_Resolved, Length: 68, dtype: float64
Cell Volume
0     946.585950
1     736.946363
2     725.419572
3     247.672352
4     602.927039
5     398.649479
6     293.973928
7     299.382990
8     319.197405
9      55.661606
10    270.130153
11    369.089344
Name: Cell_Volume_Voxel_Resolved, dtype: float64
Void Volume
0    44054.478557
Name: Void_Volume_Voxel_Resolved, dtype: float64
Move On to Next Image
Images from folder
18    MS102-4_Python
Name: IMAGE_ID, dtype: object
shape of image Z,X, Y:
(18, 1024, 1024)
Number of unique cells:
[0 2 3 4 5 6 7 8]

Impaired maturation of wild-type superoxide dismutase 1 is associated with neurodegeneration in Parkinson disease brain and a novel murine model¶

Image analysis workflow¶

The folder "IMAGES" contains the imaging data. For each replicate there is the nd2 file, a coresponding "cell" mask and a "protein" mask.¶

Create empty folders to hold information as the loop runs¶

Run the image processing steps¶